3 Kordofanian
Total Page:16
File Type:pdf, Size:1020Kb
3 Kordofanian Thilo C. Schadeberg Kordofanian languages are spoken in the Nuba Mountains in the Republic of the Sudan. The name 'Kordofanian' is an unfortunate choice since in the Sudan, particularly in the Nuba Mountains, the geographical and political term 'Kordofan' refers to the area north of the Nuba Mountains around El Obeid. It might be désirable to coin a more suitable and locally more acceptable name for this language family. The present survey chapter is based on literature and data col- lected by myself in 1974-75. 3.1 EARLY SCHOLARSHIP In 1838 Duke Maximilian in Bavaria bought the freedom of four men on the slave market in Alexandria. Taking them home, he put them in the care and under the instruction of his former tutor, Karl Tutscheck, who feit a language could be studied with the help of a single speaker even at gréât distance from that person's linguistic community. One of the four young Africans was Djalo Djondan Are from Tumale, and part of the information he provided on his people and language was published be- tween 1847 and 1853 by Karl's brother Lorenz Tutschek. Prior to this, only two short Kordofanian word lists had been published, one on 'Takele' and one on 'Schabun' (Rüppell 1829). Nothing eise followed until the end of the 19th Century except another small word list from 'Tegelé' (Munzinger 1864). A füll description of this language by Munzinger came into the possession of Reinisch at Vienna but bas since been destroyed. The fact that the Tegali dialect cluster figures in each of the early linguistic con- tacts with Kordofanian might be explained by the relative importance of the Tegali kingdom in the 19th Century. The signifïcance of this area for African language classification was realized by Lepsius (1880). In this Century several linguistic surveys have been carried out in the Nuba Mountains (Seligman 1910/11; Meinhof 1915-1919; MacDiarmid and MacDiarmid 1931; Stevenson 1956-1957; Schadeberg 1981a, b). Gradually, the extent of the Kordofanian language 68 Schadeberg family, its grammatical and lexical characteristics, and ils internai rela- tionships have become apparent. Nevertheless, the quality of this survey chapter is severely limited by the scarcity of in-depth descriptions. Meinhof (1943/44) wrote the first short grammar of any Kordofanian language, a sketch of Ebang (Heiban) based on analysis of published translations of two Gospels. The most complete existing description of any Kordofanian language is a grammar of Moro by Black and Black (1971). Fr. Carlo Muratori's notes on Orig, another dialect of the Tegali cluster, were reanalyzed and edited by Schadeberg and Elias (1979). As for the Kadugli language group (not hère considered to be Kordofanian), Reh (1985) provides an extensive and detailed description of Krongo. 3.2 CLASSIFICATION 3.2.1 INTERNAL CLASSIFICATION AND LANGUAGE INVENTORY There is one level of classification within Kordofanian that is unambiguous and non-controversial. I refer to units at this level as 'groups', and there are four such groups: 1. HEIBAN (Koalib; Koalib-Moro) 2. TALODI (Talodi; Talodi-Masakin and Lafofa) 3. RASHAD (Tegali; Tegali-Tagoi) 4. KATLA (Katla; Katla) I have added in parentheses the group names first as used by Greenberg (1963) and then those of Tucker and Bryan (1956). The group names I propose are the names of towns centrally located in the respec- tive areas; the riverine terminology populär in African linguistics cannot be applied to the Nuba Mountains for lack of permanent surface water. No confusion should arise between group names and language names since none of the place names are identical with the name any particular linguistic grouping gives itself. The overall division into groups was first established by MacDiarmid and MacDiarmid (1931). The only amendment to their grouping was Greenberg's (1950) inclusion of Tegem (Lafofa) in the Talodi group. There is a fifth group which has been classified as Kordofanian by Greenberg since 1950, i.e., KADUGLI (Tumtum; Kadugli-Krongo). I propose to exclude Kadugli from Kordofanian (see sec. 3.2.3). • HEIBAN-Group • TALOOI - Group T RASHAD-Group • KATLA - Group * KAOUGLI-Group The distribution of Kordofanian tanguages • c f c »•««»•'• »«o (Sou t h «f n Kordofon Provinc«. Sudan) PLATE 3 Kordofanian languages 70 Schadeberg TALODI group: (speakers) Ngile (Daloka, Masakin) l 12,000 Dengebu (Reikha, Masakin) -' Tocho (Moro Hills settlements) 3,800 Jomang (Talodi) 1,500 Nding (Eliri) 500 Tegem (Lafofa) 3,000 20,800 20% 40% 60% NT 80% | 1=^\ \ Ngile Dengebu Tocho Jomang Nding Tegem Tegem is very different from the other languages of the Talodi group but there is good lexical, phonological, and morphological évidence for joining Tegem and 'Narrow Talodi' into one group or branch of Kordofanian. I cannot exclude the possibility that some alleged dialects of Masakin and of the cluster called 'Moro Hills' may turn out to be sepa- rate languages. RASHAD group: (speakers) Noun dass dialects (Tagoi cluster): Goy (Tagoi), Orig (Turjok), Umale (Tumale), etc. 11,000 Classless dialects (Tegali cluster): Gom (Rashad), Tegali, etc. 25,000 36,000 A breakdown into dialects with and without noun classes appears to be thé basic genetic division of the Rashad group. The two dialect clusters share about 70% basic vocabulary (100-word list). There are a number of outlying hills about which no linguistic data is available, but where closely related dialects are supposed to be spoken. KATLA group: (speakers) Kalak (Katla) 10,900i Lomorik (Tima) 1,100? 12,000? Kordofanian 7l The two Katla languages share about 50% basic vocabulary (100- word list). Even less is known about this group than about the others. It is not clear at present whether thé four groups should be re- garded as primary branches of Kordofanian or whether intermediate lev- els of relationship should be recognized. Tucker and Bryan originally (1956) posited a 'Larger Unit' consisting of the three groups Heiban, Talodi (and their 'Single Unit' Lafofa), and Rashad. Later (1966), they separated Rashad from this 'Larger Unit' and also drew attention to Stevenson's (1957:51) observation that spécifie lexical resemblances exist between Rashad and Katla. At least since Meinhof (1915-19) thé question of historical sub- grouping and intergroup relationships has been complicated by confusing it with typological classification into languages with and without noun classes. 3.2.2 EXTERNAL CLASSIFICATION WITHIN NIGER-CONGO One of Greenberg's 1963 revisions of his own earlier classification of African languages was to join Niger-Congo and Kordofanian äs two primary branches of a single family he called Niger-Kordofanian (also: Congo-Kordofanian). Since then, the genetic relatedness between Kordofanian and Niger-Congo has not been seriously questioned. Still, some linguists believe this kind of historical relation to be too remote to be recoverable by conventional methods of historical linguistics (e.g., Möhlig 1983:175). Convincing lexical resemblances between Kordofanian and other branches of Niger-Congo are scarce. Though Greenberg (1963) lists forty-five sets of possible cognâtes (not counting seven sets linking Niger-Congo exclusively with Kadugli), it remains to be seen which and (less importantly) how many of these will survive the test of internai Kordofanian and Niger-Congo reconstruction. Bennett and Sterk (1977) concluded that their lexical évidence was insufficient for proving genetic relationship between Niger-Congo and Kordofanian (and Mande). Not wishing to reject such links altogether, they assumed an initial three-way split; this position is maintained in the present volume (Niger-Congo Overview). While the extent and importance of lexical évidence is still un- certain, there is other évidence for which the only possible explanation seems to be common historical origin: similar noun dass Systems appear in most major groupings of Niger-Congo and Kordofanian. At first sight, this may appear to be no more than a typological similarity, and as such, irrelevant to the question of genetic relatedness. However, it has to be 72 Schadeberg pointed out that this type of noun dass System is, if not unique, extremely rare in the languages of the world; the noun dass Systems found in Caucasian and Australian languages are quite different. It is also rare for a language to acquire a noun dass System through borrowing, Mbugu being the only such case ever cited. It is therefore justified to view the présence of this particular type of noun class System as an argument for genetic relations within Niger-Congo, as indeed Greenberg did (1950:393): "The very fact of the existence of noun affix classes for singular and plu- ral in Niger-Congo and Kordofanian is indicative of possible further con- nections, since such formations are rare elsewhere in thé world." At this time, Greenberg did not yet claim Niger-Congo and Kordofanian to be related. Later (1972:190) hè wrote: "It is not too much to state that in at least one instance, that of Niger-Kordofanian, grammatical resemblances were treated as more décisive than lexical ones." However, similarities between the Niger-Congo and Kordofanian noun class Systems are not only typological but can be extended to proper sound-meaning correspondences as well. The meaning of gender is to be understood not only as the vague common semantic feature of part of the nouns occurring in it, but also as that set of nouns (i.e., their meanings) itself. In the figure below (adapted from Schadeberg 1981e), six noun classes of Kordofanian are compared with what I believe are cognate forms in Atlantic, Kwa, Benue-Congo, Gur, and Adamawa-Ubangi. 1 3 4 5 6 6a plof 3 plof 5 Kordofanian gu- gu- j- H- gu- B- people tree egg liquids Atlantic go- gu- Ci- de- ga- ma- (Doneux 1975) people tree names head, name liquids Oti-Volta n -bo -Ci -djt -a -ma (Manessy 1975) people tree egg, head liquids Togo Remnant 0- o- i- U- a- N- (Heine 1968) people firewood egg, head, name liquids Benue-Congo u- u- (t)i- li- a- ma- (De Wolf 1971) people tree egg, head, name liquids Bantu NP mu- mu- mi- j- ma- nia- (Meeussen PP ju- gu- gi- li- ga- ga- 1967) people tree egg, name liquids Mba (Ubangi) -V -V -e -le -me (Carrington Num g- w- 0- 1- 1949) man name liquids FIGURE 3.1.